多媒體流量特性分析.pdf

1、西南交通大學(xué)碩士學(xué)位論文多媒體流量特性分析姓名：田志超申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：計(jì)算機(jī)應(yīng)用技術(shù)指導(dǎo)教師：譚獻(xiàn)海20090501ABSTRACTWiththeincreasinglyextensionofthemultimediaapplicationthevideoencodedwithMPEG4andH263isexpectedtoaccountforalargeportionofthetrafficinfuturenetworksSo

2、，characteristicsanalysisandmodelingofmultimediatraffichavebeenbecominglotsofresearchonnetwork艦伍CaresearchhotspotatpresentCurrentlycharacteristicsiScentralizedonthecharacterisitcatlargetimescaleConverselythecharacteristic

3、sofmultidediatrafficatsmalltimescaleisn’tinvestigatedclearlyWaveletanalysis，calledmathematicalmicroscope，isanewsubjectwhichisdevelopedonthebasisofappliedmathematicsandhasmoreadvantagescomparingtothetraditionalFouriertran

4、sformandaddingwindowFouriertransformWaveletanalysisiswidelyusedinmanybasicsubjects，especiallyinsignalprocessingandinformationscienceThisthesismadearesearchoncharacteristicsanalysisofmultimediatrafficinsmalltimescaleusing

5、WaveletanalysisThethesisfLrstsummariesthewaveletbasedjointestimatorstudieswavelettransformdeeplyandgivestheestimationprocedureandtheprincipleofmeprocedureAfteranalyzingthevanishingmomentseffectontheiointestimatoranewmeth

6、odcalledVMMSE(VanishingMomentMeanSquaredError)isputforwardwhichcangiveahelpforchoosingvanishingmomentstominimizetheinfluenceofvanishingmomentonwaveletanalysisSecond，thestabilityandGaussianpropertyofLANandWANtra墑cinsmallt

7、imescaleisinvestigatedTheresultsofourresearchshowthatthetrafficisnotofstabilityandGaussianinsmalltimescaleAmethodcalledRunTest，whichisaquantitativemethodtotestthestabilityoftraffic，isputforwardThismethodgivesUSagoodwayto

8、teststabilityFinallythevideosencodedwithMPEG4andH263aleanalyzedsystematicallyThethesismainlyinvestigatedfirstmoments，longrangedependentfractalandmultiscalepropertiesMeanwhile，theinfluenceofcodingqualitycodingrateandvideo